Keeping It in the Family: HER3 as a Target in Brain Metastases

In this issue of Clinical Cancer Research, Tomasich and colleagues report that HER3 is expressed in brain metastases (BM) from patients with breast cancer and non–small cell lung cancer (NSCLC; ref. 1). They compare expression of HER2 and HER3 showing that HER3 expression is maintained in tumors that are HER2 negative by IHC (0). They show that 75% of breast cancer BMs (BCBM) and 72.9% of NSCLC BM express HER3. When comparing matched NSCLC primary and BM tissues, they show that BMs are enriched in HER3 expression (72.3% vs. 40.4%). To explore downstream pathways activated by HER3 expression, they performed methylation profiling in HER3-positive and HER3-negative BM samples (16 NSCLC and 55 breast cancer) to identify pathways that were enriched between breast cancer subtypes and some that were common such as PI3K-AKT signaling, BDNF-Trk signaling, and MAPK signaling. HER3 expression was associated with improved overall survival in patients with estrogen receptor (ER)-positive HER2-negative BM; however, numbers in this cohort were very small, (41 HER3-positive vs. 8 HER3-negative) and patients were treated over a long time span, which could introduce additional imbalances that could confound overall survival associations.

HER3 is localized in the long arm of chromosome 12, encoding a 180 kDa protein that has little intracellular kinase activity (2). Activation of HER3 depends on heterodimerization with other HER family members, preferentially HER2 and EGFR (Fig. 1A). HER3 expression has been detected in breast, lung, ovarian, colon, gastric, cutaneous and pancreas cancers, but is not oncogenic when overexpressed alone (2).

Previous work has shown that HER3 protein expression was enriched in BCBMs compared with matched primary tumors (59% vs. 29.7%, P = 0.019), as measured by IHC (3). Similarly, a study comparing phosphorylated HER3 expression in NSCLC showed enrichment in matched BM tissue compared with primary tumors (mean score difference 21.3, P = 0.001), while a separate study found that the frequency of HER3-positive NSCLC tumors was also enriched in brain (91%) versus primary (80%) site (4, 5).

This study raises the question about the role of HER3 in the development and targeting of BMs in breast and NSCLC. The cognate ligands for HER3 are neuregulins 1 and 2. Neuregulins are EGF-like signaling molecules, expressed by microglia and neurons within the brain and other organs (6). When HER2-amplified or PIK3CA-mutant breast cancer cell lines were implanted in mice brains, compared with mammary glands, they preferentially increased expression of HER3, partly in response to paracrine and autocrine NRG-1 expression, and were resistant to PIK3CA inhibition (7). However, in the same study, targeting HER3 using pertuzumab (blocks HER2/3 dimerization) or LJM716 (an allosteric HER3 antibody) did not reduce tumor growth compared with vehicle. In the phase I trial of U3-1287 (patritumab), a HER3-targeting mAb that blocks ligand binding, there was only one partial response, with 40.4% progressive disease and 24.6% with stable disease (8). Previous work showed that the preclinical activity of HER3-targeting antibody LJM716 against BCBM could be improved by adding the HER2-targeting antibody pertuzumab, or PIK3CA inhibitor buparlisib (7). This suggests that an effective therapeutic strategy for targeting HER3 in the BCBM, and other sites, depends on augmented inhibition of pathways downstream from HER3.

Seribantumab is a fully human IgG2 mAb targeting HER3 and blocking NRG-mediated HER3 signaling (2). When tested in a randomized phase II trial in combination with erlotinib for patients with metastatic NSCLC (9), the addition of seribantumab did not improve progression-free survival compared with erlotinib alone. A similar lack of progression-free survival benefit was seen in a phase II trial of seribantumab added to exemestane in patients with metastatic ER-positive HER2-negative breast cancer (10).

Work by Tomasich and colleagues demonstrates that HER3 expression in BM from patients with metastatic breast cancer and NSCLC is associated with diverse changes in the tumor methylome (1). The heterogeneity of HER3-activated redundant and bypass mitogenic and prosurvival pathways (Fig. 1A) may explain why HER3 monotherapy, or even combinations with targeted drugs (e.g., HER3 plus PIK3CA inhibitors), have had modest clinical impact (11).

As a class, antibody–drug conjugates (ADC) have rapidly transformed the treatment landscape across multiple solid tumors. In contrast to naked mAbs, HER3-targeting ADCs carrying an anti-topoisomerase 1 payload (e.g., HER3-DXd) function as highly potent cytotoxics and effectively bypass multiple upstream mitotic signaling pathways to induce DNA damage, interrupting cell proliferation and resulting in tumor cell death (12).

Patritumab deruxtecan (HER3-DXd) has shown promising activity in patients with HER3-expressing metastatic breast cancer and NSCLC with and without an EGFR mutation (13–15). In the phase I/II trial of HER3-DXd in HER3-expressing metastatic breast cancer, the objective response rate (ORR; across all doses) was 30.1% in hormone receptor–positive HER2-negative subset, 22.6% in the triple-negative breast cancer subset and 42.9% in the HER2-positive subset. All subsets reported were also HER3 high (>75% membrane staining; ref. 13). In the phase I trial of patients with EGFR-mutant NSCLC that progressed on previous tyrosine kinase inhibitor (TKI), the ORR was 39% at 5.6 mg/kg dose. In this study, 21% (17/81) had stable BMs at baseline but progression by site was not reported (14). The authors found that responses to HER3-DXd occurred across a broad range of HER3 IHC expression scores (most were high), similar to what has been reported for anti-trophoblast cell-surface antigen 2 (anti-Trop-2) targeting ADC, sacituzumab govitecan in metastatic breast cancer (16). In a separate phase I study of patients with NSCLC without EGFR-mutation, the ORR was 28%, and patients with stable BMs are eligible (15). In the follow-up HERTHENA-Lung02 phase III trial of patients with metastatic EGFR-mutant lung cancer that have progressed on an EGFR-targeting TKI, patients with active BMs are excluded (NCT05338970).

The current study highlights a crucial opportunity to test whether HER3-targeted ADCs can exert antitumor efficacy in the central nervous system (CNS), and thereby address an area of tremendous unmet medical need in patients with breast cancer, lung cancer, and beyond. It has been accepted dogma that due to their large molecular size and purported inability to cross the intact blood–brain barrier, ADCs must be inactive in the CNS. However, multiple lines of evidence now directly contradict this assumption and strongly support a strategy of including patients with active BMs throughout all phases of ADC clinical development (17, 18). High levels of the ADC payload SN-38 were measured in resected CNS lesions in patients with breast cancer or glioblastoma multiforme treated preoperatively with sacituzumab govitecan with preliminary evidence of efficacy (19). In the KAMILLA study, an exploratory analysis found that treatment with the ADC trastuzumab emtansine (T-DM1) was associated with 49.3% ORR in patients with active BM not previously treated with radiotherapy (20). More recently, three recent studies demonstrated the intracranial activity of trastuzumab deruxtecan (T-DXd), a HER2-directed ADC with a cleavable tetrapeptide based linker and a topoisomerase1 inhibitor payload (DXd) in patients with active breast cancer BMs (21, 22). In TUXEDO-1, a single-arm phase II trial of patients with HER2-amplified untreated or progressive BCBMs treated with (T-DXd) the intracranial ORR was 73.3% (21). In DEBBRAH, a multicenter, phase II, single-arm trial of patients with HER2-positive or HER2-low BCBMs, the intracranial ORR in patients with active BMs was 6/13 (46%) (23). In a retrospective real-world cohort of patients with HER2-amplified BCBM treated with T-DXd, the subset with progressive or untreated BCBMs had an intracranial ORR of 70% (22).

While the data support the potency of systemic therapies for treatment of BMs, early and even late-phase studies of promising anticancer therapies continue to exclude patients with active BMs, despite strong recommendations from American Society of Clinical Oncology, Friends of Cancer Research, the RANO (Response Assessment in Neuro-oncology Criteria) working group, and patient advocates (24). Critically, Tomasich and colleagues add to a growing body of literature demonstrating the clinical utility of a translational workflow (Fig. 1B) that aims to target BM-specific dependencies through site-specific omic-analysis, orthotopic tumor modeling and drug testing, followed by design and reporting of BM-specific clinical trials (21, 22).

N.U. Lin reports grants from Genentech, Pfizer, and Zion Pharmaceuticals; grants and personal fees from SeaGen, Olema Pharmaceuticals, and AstraZeneca; personal fees from Stemline/Menarini, Daiichi Sankyo, Blueprint Medicines, Janssen, Prelude Therapeutics, Aleta Biopharma, Voyager Therapeutics, Denali Therapeutics, Up To Date, and Affinia Therapeutics; and non-financial support from Artera Inc outside the submitted work. No disclosures were reported by the other author.